Close Menu
    Facebook X (Twitter) Instagram
    Facebook X (Twitter) Instagram
    StockNews24StockNews24
    Subscribe
    • Shares
    • News
      • Featured Company
      • News Overview
        • Company news
        • Expert Columns
        • Germany
        • USA
        • Price movements
        • Default values
        • Small caps
        • Business
      • News Search
        • Stock News
        • CFD News
        • Foreign exchange news
        • ETF News
        • Money, Career & Lifestyle News
      • Index News
        • DAX News
        • MDAX News
        • TecDAX News
        • Dow Jones News
        • Eurostoxx News
        • NASDAQ News
        • ATX News
        • S&P 500 News
      • Other Topics
        • Private Finance News
        • Commodity News
        • Certificate News
        • Interest rate news
        • SMI News
        • Nikkei 225 News1
    • Carbon Markets
    • Raw materials
    • Funds
    • Bonds
    • Currency
    • Crypto
    • English
      • العربية
      • 简体中文
      • Nederlands
      • English
      • Français
      • Deutsch
      • Italiano
      • Português
      • Русский
      • Español
    StockNews24StockNews24
    Home » Quarterly Earnings: Signal vs. Noise, Cost vs. Benefit
    Fund News

    Quarterly Earnings: Signal vs. Noise, Cost vs. Benefit

    userBy user2025-09-18No Comments5 Mins Read
    Facebook Twitter LinkedIn Telegram Pinterest Tumblr Reddit WhatsApp Email
    Share
    Facebook Twitter LinkedIn Pinterest Email


    With the White House downplaying the value of quarterly reporting for companies, investors face a familiar question: does the cost of producing information outweigh the benefits?

    Using Robert Shiller’s long-run data, this post shows that quarterly earnings contain information that is likely valuable to both long-term allocators and short-term traders. Its benefits, which I don’t attempt to quantify, should be weighed against any savings from less-frequent reporting.

    Quarterly vs. Semi-Annual: What’s at Stake

    The White House this week called for a change from quarterly to semi-annual earnings reporting. President Donald Trump argued that such a shift would save companies money and time.

    That may be true. But would investors lose valuable information?

    To answer this question, I use earnings data from Robert Shiller’s online data from January 1970 (1970:1), the year in which the Securities and Exchange Commission made quarterly earnings mandatory, to 2025:6 to test relationships among the change in three-month earnings, six-month earnings, and the trend in earnings. I define the trend as a 61-month centered moving average change in earnings. Specifically, I test whether knowing three-month earnings’ changes helps an investor better estimate changes in the longer-term trend in earnings.

    Chart 1 shows three-month earnings in green, six-month earnings in red, and trend earnings in blue. Series start in January 2000 (2000:1), rather than 1970:1, for ease of visualization.

    Chart 1. 3-month, 6-month, and trend earnings, 2000:1 to 2025:6.

    Source: Robert Shiller online data, author calculations.

    Of course, three-month earnings are choppier than six-month earnings. But it is not obvious from visual inspection that knowing three-month earnings in addition to six-month earnings would help a long-term investor predict changes in trend earnings. (I test this below and find that they may).

    It is, however, obvious that a short-term investor, one perhaps interested in earnings changes in periods of less than a year, would benefit from knowing three-month earnings. This observation is confirmed empirically below.

    subscribe

    I start with the long-term investor, who I assume is interested in the long-term trend in earnings. A natural way to gauge the value of having three-month earnings in addition to (or instead of) six-month earnings is to model the change in trend earnings as a function of one or both, estimate that model using ordinary least squares, and compare model accuracy. In this post, I use R-squared as my measure of fit (or adjusted R-squared) — the larger, the better.

    At any point, the investor knows one-half the current trend in earnings. That is, they know the first 30 months’ earnings of the current 61-month window, my proxy for the trend in earnings. And they know either the last three months of earnings, or the last six months of earnings, or both.

    To determine whether receiving earnings information every three months as opposed to every six months would help the long-term investor to better predict the trend, I estimated specifications where the change in 30-month-ahead trend inflation is explained by the change in six-month earnings alone plus the prior earnings-trend change (Model 1). In Model 2, the trend change is explained by the same variables plus the three-month change in earnings. Results are shown in Table 1.

    Table 1. Regressions of trend inflation change on 3- and 6-month earnings changes, 1970:1 – 2025:6.

    Dependent variable = Trend inflation (30-month lead)
      Model 1 Model 2
    Six-mo. change (three-mo. lag) 0.073 (0.013) 0.061 (0.013)
    Three-mo. change – 0.124 (0.029)
    Trend change -0.223 (0.041) -0.234 (.040)
    Adjusted R-squared 0.098 0.126
    Obs 547 547

    Source: Robert Shiller online data, author calculations.

    Since I’m not interested in inference, I omit discussion of estimated coefficient values, other than to note that they enter with the expected sign. Notwithstanding this, I include the prior trend in earnings to reduce bias in my estimates and standard errors appear in parenthesis next to each estimate.

    The key result is that adding quarterly earnings (three-month change) improves fit — the adjusted R-squared increases from 0.098 for Model 1 to 0.126 for Model 2. While neither fit is impressive, these results suggest that quarterly earnings may help the long-term investor predict trend earnings. Other measures of fit, namely the Akaike and Bayesian information criteria (AIC and BIC), confirm that the specification which includes 3-month earnings is more accurate.

    As for what may be of interest to traders (short-term investors), one might guess that the three-month earnings change is related to the next three-month change. Quarterly earnings changes are indeed persistent. The scatter in Chart 2 shows the autocorrelation of quarterly earnings, where extreme values (earnings changes greater than 100%) have been removed for easier viewing. The estimated slope is 0.601 (se = 0.031) — the blue best fit line is flatter than the black 45-degree diagonal line — and the R-squared is 0.361.

    Chart 2. Three-month lagged earnings change vs. three-month earnings change, 1970:1 – 2025:6.

    Source: Robert Shiller online data, author calculations.

    And at the risk of estimating the obvious, the R-squared of a model explaining 12-month earnings with six-month earnings (from six-months before) is 0.699, whereas including three-month earnings (from three-months before) improves the fit to 0.953.

    Cost vs. Benefit

    It is nearly axiomatic that, in most applications, more data is preferable to less. And the results discussed here suggest that quarterly earnings contain valuable information for investors. But producing earnings is costly.

    As regulators consider reducing reporting frequency, they should weigh not just the savings but also the potential losses — losses to investors resulting from less transparency and to the economy resulting from impaired market efficiency.

    More to Think About

    Past CFA Institute member surveys show clear support for quarterly earnings.




    Source link

    Share this:

    • Click to share on Facebook (Opens in new window) Facebook
    • Click to share on X (Opens in new window) X

    Like this:

    Like Loading...

    Related

    Share. Facebook Twitter Pinterest LinkedIn Tumblr Telegram Email
    Previous ArticleSolar power cuts electricity bills and carbon emissions—NZ needs to scale up faster
    Next Article Private sector encouraged to avail of carbon credits
    user
    • Website

    Related Posts

    When the Fed Cuts: Lessons from Past Cycles for Investors

    2025-09-17

    Coaching Investors Beyond Risk Profiling: Overcoming Emotional Biases

    2025-09-16

    How to use the Freefincal Equity Mutual Fund Screener

    2025-09-16
    Add A Comment

    Leave a ReplyCancel reply

    © 2025 StockNews24. Designed by Sujon.

    Type above and press Enter to search. Press Esc to cancel.

    %d